Estimating oxygen exchange across the air–water interface of a hypereutrophic lake

Abstract

Patterns of estimates of oxygen flux (J) across the air–water interface of hypereutrophic Onondaga Lake, NY, U.S.A., are characterized for time scales ranging from diel to seasonal for an 8-month period. The analysis is supported by a high frequency (most often hourly) monitoring program, conducted with a robotic buoy, that included measurements of dissolved oxygen (DO), temperature, and fluorometric chlorophyll a in the lake's surface waters, vertical profiles of DO through the epilimnion, and wind speed and solar radiance. The magnitude and direction of J is demonstrated to vary dramatically at diel, day-to-day, and seasonal time scales. Thus, large errors in estimates of J may result from extrapolating flux calculations made from short-term data to longer time periods. The variations in J were driven by variations in metabolic activity and meteorology, and were mediated by departures from equilibrium DO concentrations and wind-driven turbulence. Extended periods of high J values are shown to coincide with intervals of large departures from equilibrium DO concentrations, but day-to-day differences are driven mostly by variations in wind. A distinct diel pattern of J estimates is manifested for average conditions, with substantially higher J values during daylight hours. This pattern reflects the common diel patterns of the drivers of both higher DO oversaturation and wind speed over those hours. It is demonstrated that the magnitude of J is substantial relative to net changes in the epilimnetic DO pool, and thus must be accommodated accurately in estimates of primary production and community respiration that are to be based on diel monitoring of DO in the water columns of productive lakes.